基于最大似然分类算法的土地利用变化特征分析  

Analysis of Land Use Change Characteristics in Urumqi City Based on Maximum Likelihood Classification Algorithm

作　　者：李小玉 蒲智[1] 李全胜[1] LI Xiaoyu;PU Zhi;LI Quansheng(School of Computer and Information Engineering,Xinjiang Agricultural University,Urumqi 830052)

机构地区：[1]新疆农业大学计算机与信息工程学院,乌鲁木齐830052

出　　处：《计算机与数字工程》2022年第6期1153-1157,1201,共6页Computer & Digital Engineering

基　　金：国家自然科学基金项目“新疆林木腐烂病风险时空分析及预测研究”(编号:32060321)资助。

摘　　要：随着计算机技术的发展遥感影像的分类方法也逐渐增多,传统的目视解译方法逐渐被各类分类算法替代,目前基于ENVI软件的分类算法已逐渐趋于成熟,论文选取ENVI中的最大似然分类算法,以乌鲁木齐市为研究区域,选取2000年、2010年、2019年的三期Landsat影像为数据源,验证算法分类精度的同时分析得出研究区域的土地利用变化特征。结果表明:最大似然法在研究区内的总体分类精度达到87.93%,其kappa系数为0.78。乌鲁木齐市的用地类型以草地和未利用地为主,其次是耕地、林地和建设用地,水域面积所占比最小。近20年中建设用地变化最为明显,一直呈现增长趋势,于初期相比其占比增长了3.56%。With the development of computer technology of remote sensing image classification method also gradually increased,the traditional visual interpretation method is gradually replaced by all kinds of classification algorithm,the classification algorithm based on ENVI software has been gradually mature,this article selects the maximum likelihood classification algorithm in ENVI,in Urumqi city as the study area,2000,2010,2019,three Landsat images as the data source,verify the algorithm of classification accuracy at the same time analysis of land use change characteristics of the study area.The results show that the overall classification accuracy of the maximum likelihood method in the study area reaches 87.93%,and its Kappa coefficient is 0.78.Grassland and unused land are the main land types in Urumqi,followed by cultivated land,woodland and construction land,and water area accounted for the smallest proportion.Construction land has changed most obviously in the past 20 years,showing a growing trend.Compared with the initial period,the proportion of construction land has increased by 3.56%.

关 键 词：最大似然分类法 土地利用变化 乌鲁木齐市 分类精度 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置] F321.1[自动化与计算机技术—控制科学与工程]